The Size of Thoughts

by Nicholson Baker

  But no matter how clever and successful OCLC’s new quality-control efforts are, they mainly benefit those libraries that have not yet converted their catalogs. Countless old errors and inconsistencies are out there still, doing indolent mischief to scholarship in the local online catalogs of the libraries that “reconned” early on. Having paid millions in fees to OCLC for the use of its database, university libraries must now scrape up the cash to pay for authority-control software just so that their online catalog will perform the minimal tasks that Charles Cutter expected of the card catalog as a matter of course. The University of Chicago, even as it pays OCLC’s RETROCON department to convert cards relating to the classics, philosophy, and American literature (at a cost of around two dollars a card), is contemplating paying Blackwell North America, Inc., a database processor, at least a hundred and fifty thousand dollars for a onetime authority-control grooming. (Errors and inconsistencies that appear after Blackwell is finished will persevere, of course.) The CARL Corporation, in Denver, Colorado, charges in the six figures to license its authority-control application to a major university library. This can all suddenly seem very unfuturistic and sad—sad because the cost of technology now consumes nearly 30 percent of the typical American library’s budget, according to one 1992 estimate, forcing it to cut book purchases, reference staff, and skilled catalogers, and sad because the technology that libraries are actually buying turns out to be remedial software meant to correct the hash that earlier technologies have made of information once safely stored on paper.

  What we have already begun seeing, in fact, especially at state universities with dwindling budgets, is a kind of self-inflicted online hell, in which the libraries are forced to continue to pay paraprofessionals to convert their huge card catalogs, since they’ve already pillaged the paper database to the point where its integrity is unrestorable, and yet they aren’t able to afford the continuous hardware and software upgrades necessary to make the growing mass of online records function together adequately. They can’t go back, and they don’t have the money to go forward.

  I’m thinking, for instance, of U.C. Berkeley. Berkeley has one of the best research collections in the world, and the quality of its cataloging over the past hundred and twenty-five years has been unusually high. Thus its card catalog and its shelf list were filled with richly detailed, accurate, intelligent cards, many of which were sent to Maureen Finn’s RETROCON staff for conversion, sent back, and scrapped. But higher education is in serious trouble in California, and, as a result, Berkeley has chosen not to pay Maureen Finn for a premium RETROCON job. For about two dollars a card, Berkeley is getting a middling RETROCON—better than Harvard’s but still very plain. Typically, an OCLC operator takes a Berkeley card, finds a match in the database as fast as he or she can, and accepts what the database offers, without being able to spend additional time entering the supplemental (or superior) information—the notes, the subject tracings, the holdings records—that the original card may contain. “The standards of conversion were, of necessity, because of lack of funds and lack of staff, not exquisitely high,” one Berkeley employee told me. Because it can’t afford better, Berkeley (like Harvard, like hundreds of other libraries) is paying OCLC not to improve but to denature and often to mediocritize its records of old and out-of-print material. “No matter how good it is,” the employee said, “you throw away that card and, in some cases, accept something that’s inferior.”

  Berkeley’s subject catalog is already gone. Though it was more or less frozen in the 1980s, it was nonetheless very efficient for some kinds of searches: if the primary sources in your field were published in the eighteenth and nineteenth centuries, you don’t want your retrievals spread among the thousands of irrelevant records that a mechanical database subject haul brings up. Any drawer of an out-of-date paper catalog represents the equivalent of a filtered computer search, a Boolean date-limited search, of a very sophisticated sort—in fact, one drawer represents the outcome of a kind of search that most online catalogs can’t and won’t ever be able to perform, since it offers clues to what books were in the library during different eras. If in seventy years a historian of science (say) wants to know whether some Nobel-laureate physics professor could possibly have seen and been influenced by a certain out-of-the-way Dutch mathematical monograph from the thirties that bears important similarities to the professor’s work—whether, that is, it was part of the library’s collection during the period when the professor was developing his ideas, or was acquired only after the professor’s papers were published (perhaps acquired by the library as a gift from the professor’s estate, because the professor was sent the monograph by the Dutchman himself, anxious to establish primacy)—the historian of science will have little chance of finding an answer to his question now, because the computer record will bear the (to him) meaningless date of the retrospective conversion of the card, i.e., sometime in the late eighties or early nineties, which has no relation to the time that the card was originally produced (and the book placed on the shelf), whereas the original card, even if it bore no direct date of creation, would have exhibited distinct features (typewriter style, format, cataloging conventions) that might have enabled a catalog-card paleographer to place it within a five-year period.

  Why, then, did library administrators order the Berkeley subject catalog destroyed? Was it really just to have the space for eight study tables? Admittedly, eight study tables, incised with the inevitable obscene drawings, declarations of love, and reciprocal ethnic slurs, and populated with thirty or forty pre-midterm Psych 101 students making soft sighing noises with their pungent highlighters and burping like moss-gorged moose from time to time for comic effect, is a noble sight to have in a library. But is it a reasonable trade-off? Library administrators always use the magical phrase “out of space” when they want to get rid of something, but this in no way constitutes an argument. Libraries have been running out of space since the Sumerians first impassioned clay, because tablets and scrolls and manuscripts and books and microforms and computer disks tend to take up space, and their numbers inevitably grow. There are countless duplicates of old textbooks from the sixties and seventies on the shelves of most university libraries; scholarly and scientific journals probably pump more non-unique paper into Berkeley’s library system every few months than was contained in its unique subject catalog. A library continues to buy books, and it selects what it throws out, on the basis of what it judges is of value to present and future users of the library: the need for space is merely a constant, SNEED, in every decision to acquire or discard.

  Administrators are singling out card catalogs, I think, not as a last resort but as a first resort, because they hate them. They feel cleaner, lighter, healthier, more polyunsaturated, when all that thick, butter-colored paper is gone. (One Berkeley administrator was heard saying, “Oh, we’re just going to get rid of those junky old cards. Nobody uses the subject catalog anyway.”) “Resist the impulse to burn those old cards,” gaily cautions one article on retrospective conversion, since “the staff will experience withdrawal symptoms.” The impulse to burn is there, it seems to me, because library administrators (more often male than female) want so keenly to distance themselves from the quasi-clerical associations that surround traditional librarianship—the filing, the typing, the shelving, the pasting, the labeling. Librarianship, they think (rightly), hasn’t received the respect it deserves. The card catalog is to them a monument, not to intergenerational intellect, but to the idea of the lowly, meek-and-mild public librarian as she exists in the popular mind. The archetype, though they know it to be cheap and false, shames them; they believe that if they are disburdened of all that soiled cardboard, they will be able to define themselves as Brokers of Information and Off-Site Digital Retrievalists instead of as shy, bookish people with due-date stamps and wooden drawers to hold the nickel-and-dime overdue fines, with “Read to Your Child” posters over their heads and “February Is Black History Month” bookmarks at th
eir fingertips. The proponents of computerization are such upbeat boosters of the library’s potential role in the paperless society (Fred Kilgour once wrote that “not having to go to a library is a very important improvement in providing library service,” and when asked in an end-of-career interview whether he felt there was any possible downside to library automation, he thought for a moment and replied, “I can’t think of any negative effect”) that library managers are encouraged to forget—are eventually frightened even to admit—that their principal job is to keep millions of used books dry and lend them out to people. When we redefine libraries as means rather than as places—as conduits of knowledge rather than as physical buildings filled with physical books—we may think that the new, more “visionary,” more megatrendy definition embraces the old, but in fact it doesn’t: the removal of the concrete word “books” from the library’s statement of purpose is exactly the act that allows misguided administrators to work out their hostility toward printed history while the rest of us sleep.

  Again, lest we become confused and forgetful, the function of a great library is to sort and store obscure books. This is above all the task we want libraries to perform: to hold on to books that we don’t want enough to own, books of very limited appeal, unshielded by racks of Cliffs Notes or ubiquitous citations or simple notoriety. A book whose presence you crave at your bedside or whose referential or snob value you think you will need throughout life, you buy. Libraries are repositories for the out of print and the less desired, and we value them inestimably for that. The fact that most library books seldom circulate is part of the mystery and power of libraries. The books are there, waiting from age to age until their moment comes. And in the case of any given book, its moment may never come—but we have no way of predicting that, since we are unable to know now what a future time will find of interest.

  So the incremental value of any one library book, even a rare and costly book, is tiny. The value of a huge collection isn’t immeasurably increased by the acquisition of some out-of-date 1943 monograph on, say, granary design. And, conversely, the loss of any one book does no obvious harm to the whole. If an operator reverses two numbers in transferring the call number for a book from a card, causing you to look for the book on a shelf where it isn’t, how terrible is that? The numerical typos and coding errors in online catalogs that (particularly in our era of closed, unbrowsable stacks) result in the complete disappearance of a title for the seeker—in its effective, though not physical, loss, its total unfindability, its sinking from view—can’t compel outrage, or make headlines in academia, because most titles are in the minds of no more than five or ten people at a time. Sometimes the only person who has devoted fifteen minutes of mental receptivity and appreciation to a book, aside from its author and (with luck) its publisher, is the cataloger at the library who described it for the OCLC database. If a hundred thousand volumes disappeared at random from the shelves of a major university library in a single night, it might take weeks or months before graduate students compared their puzzling shelf experiences and slowly realized that something big, something on the order of Fahrenheit 451, had taken place. And yet, despite the insignificance of any individual wheat stalk of a book in relation to the total Nebraska of print, we demand that libraries exercise extraordinary care to preserve the ephemera they have on their shelves; we want them to get right to work microfilming or digitizing whatever starts crumbling; we want to trust that when we pass through a university library’s entrance turnstile we aren’t going to be missing too much of the inspiringly miscellaneous assembly of all that has been done and thought.

  We certainly don’t think that because a book may wait a decade or three between checkouts the library should necessarily cull it. And even if we do consent to the culling, and forgive the library for laying it out on the twenty-five-cent table, we are unlikely to agree that all copies of that book, in all libraries, private and public, ought to be rooted out and destroyed. We want the book to continue to exist somewhere, not to go extinct, because in some later ecosystem of knowledge it may be put to some surprising use—a cautionary use, a comic use, a cultural-historical use. And unforeseen secondary uses await the book that every library is certain to have self-published, as well. By studying Haverford’s card catalog, Michael Stuart Freeman, librarian of Haverford College, was able to determine when his predecessors deployed their first typewriter (it was during the summer of 1916)—a tiny fact, perhaps, but one of sufficient interest to Freeman that he published a brief, thoughtful paper that discussed the history of typewriters in libraries. Freeman kept several handwritten cards as samples (“I save a few, because I’m a sentimentalist,” he confided to me); but once his own passing curiosity was satisfied, he dumped the rest of his catalog.

  Put in mind of mass extinctions and systems analysis, I got in touch with Jim Bradley, a programmer-analyst at the Computer Services department of Berkeley’s College of Natural Resources and the author of an unfinished entomological dissertation entitled “Computer Tools for Pest Management: A Case Study of the Codling Moth.” I asked him about retrospective conversion. “At universities, people get financial support if they’re doing something sexy, and an awful lot of people that are running the over-all library operation today are not so much reference librarians as promoters of sexy modern technique,” he promptly said. “The people who pay the bills want to get out of the stuff business. They don’t want libraries to have anything in them.” I fulminated ineffectually about the new typo layer in online catalogs, and he replied by making a point that seems to me undeniable: “I guess what we’re really doing is we’re having a short Dark Ages of scribalism as we transcribe from the original records into the electronic form. There’s going to be that same blot on the historical record in our age as there was in the Middle Ages.”

  We should know better than to do this to ourselves. Or, if we do do this to ourselves—make a gigantic software upgrade of sorts from a paper database to an electronic one, because it’s inevitable—we should, good systems managers that we are, have the sense to keep the old “software” around as a backup, in case Rev. 2.0 has strange bugs and doesn’t perform as claimed. Charles Hildreth, a big name in library automation, said in a 1985 interview that “comparing the … card catalog to the online catalog is like comparing a bicycle to a space vehicle; they’re both modes of transportation but that’s where the similarities end.” And the question is: Don’t both modes have characteristic and complementary virtues? Which mode do you really want to ride to school? Which is going to have the brittle O-rings? Which is costlier? Which is likelier to drift aimlessly off into outer darkness? Which would you prefer to have your fourth grader ride? In a study of some fourth, sixth, and eighth graders carried out at the Downers Grove Public Library, in Illinois, Leslie Edmonds found that 65 percent of the kids’ card-catalog searches were successful, versus only 18 percent of their online searches. No fourth graders used the online catalog successfully. (They were asked to look up things like “Fire Stations,” “Insects—Poetry,” “Octopus Pie,” and “The Curse of the Blue Figurine.”)

  And grown-ups have problems, too. There are many more ways to go astray if you must type your way to the call number for a book than if you can flip mutely through cards to one. In 1984, Jean Dickson, in An Analysis of User Errors in Searching on Online Catalog, found that users of Northwestern’s LUIS failed 39.5 percent of the time to type in a title at a terminal in a way that would bring up the record for something that actually existed in the database. She mentions, as one class of keyboard entries, “expressions of frustration” such as “Bleahh!” and “I hate this computer!” and various obscenities.

  Under the “Bleahh!” subject heading may be adduced the Problem of Accelerating Typos. First, you type in a name and get no record. You think, Is it me, or is it that the library doesn’t have anything by this person? You look at what you’ve typed. No obvious mistakes. You try again, several ways. Could this library possibly not have anythin
g by this person? You have to decide whether it’s a question of your having given a bad command, or a matter of a harder-to-spot typo, or a variant form of the writer’s name. Some online catalogs require “A” as the author-search command, some “A=,” some “FI PA” (for “Personal Author”), some “FI PN” (for “Personal Name”), some “FI AU,” some “AUT” or a number from a menu. Some screens respond as soon as you’ve typed the magical letter, some want you to press Return to deliver the command. It’s as if you walked up to a card catalog you hadn’t used in a while and weren’t sure whether, in order to open a drawer, you were supposed to pull on the drawer handle, push on the drawer handle, twirl the brass end of the holding rod, or fart twice and sing “God Bless America” in a hoarse falsetto. Anger builds. You’ve forgotten whether you’ve already tried some variant, because your previous tries have disappeared off the screen. With anger comes poor typing. You can’t know it, but you have finally found the proper command and the right form of the name—it’s just that now you’re so steamed that you’re making basic typos every time. Finally everything gels, and you get 123 RECORDS RETRIEVED BY YOUR SEARCH, and then you learn, to your further dismay, how extremely long it takes to page through that seemingly small number of records (“at the simple touch of a button”), hitting Enter, Enter, Enter, or M, M, M, or F, F, F.